Loading ...
Sorry, an error occurred while loading the content.

Q: CentOS/RHEL - what is it that puts up the graphic login window

Expand Messages
  • thad_floryan
    For 6 weeks everything s fine with the new system, but now, beginning exactly on May 1, the login window (for username/password) takes many minutes before it
    Message 1 of 11 , May 1, 2012
    • 0 Attachment
      For 6 weeks everything's fine with the new system, but now, beginning
      exactly on May 1, the login window (for username/password) takes many
      minutes before it appears. Is this some kind of a statement by RHEL
      concerning Soviet Russia's "May Day" celebrations or what? :-)

      I can ssh into the system just fine within 20 seconds or so of powering
      up, but the graphic login screen is delayed and there is no disk
      activity whatsoever for quite a long time. I didn't time the number of
      minutes. Once logged on, I can logout and a new login window appears
      almost instantly, so this is something related to being just-booted.

      Is that login window something from Pam and/or gdm-password? Googling
      isn't helping [me] find what it's called so I can seek a solution.

      grep'ing in /var/log for gdm and other things isn't finding anything
      useful, neither is dmesg (both the program and grep'ing the file).

      Nothing obvious is in /var/log/messages either (perhaps because I
      don't know what I'm looking for).

      If anyone could reveal the name of the process that puts up the login
      window, I could take it from there. SELINUX is *NOT* enabled as best
      as I can tell. And, FWIW, I changed NOTHING in the system config for
      at least a month now, and today, May 1, is the first instance of this
      new login problem (noting I power-down each night and reboot each
      morning).

      I also don't recall any of the almost-daily updates from CentOS having
      anything to do specifically with logins, so this is a puzzler -- what
      could have changed? Sigh.
    • Scott
      ... What happens if you boot into runlevel 3 and then do startx. I put nothing past Gnome people or RH in breaking things. ... sestatus is the command to
      Message 2 of 11 , May 1, 2012
      • 0 Attachment
        On Tue, May 01, 2012 at 10:57:28AM -0000, thad_floryan wrote:


        >
        > I can ssh into the system just fine within 20 seconds or so of powering
        > up, but the graphic login screen is delayed and there is no disk
        > activity whatsoever for quite a long time. I didn't time the number of
        > minutes. Once logged on, I can logout and a new login window appears
        > almost instantly, so this is something related to being just-booted.

        What happens if you boot into runlevel 3 and then do startx. I put
        nothing past Gnome people or RH in breaking things.
        >
        > If anyone could reveal the name of the process that puts up the login
        > window, I could take it from there. SELINUX is *NOT* enabled as best
        > as I can tell.

        sestatus is the command to check that. Also /etc/sysconfig/selinux.


        setenforce 0 will change status to off, but that will be lost with
        reboot. setenforce=0 can be added to the kernel line though, if you
        want to boot it once to check. (Note that, especially if you're
        generally using it, it's much quicker to change, in
        /etc/sysconfig/selinux, to and from permissive--if you change to and
        from disabled/enabled, it will relabel upon bootup.


        --
        Scott Robbins
        PGP keyID EB3467D6
        ( 1B48 077D 66F6 9DB0 FDC2 A409 FA54 EB34 67D6 )
        gpg --keyserver pgp.mit.edu --recv-keys EB3467D6
      • thad_floryan
        ... I haven t explicitly tried that (yet) because X is running fine, it s just the presentation of the login window is delayed many minutes. We already know
        Message 3 of 11 , May 1, 2012
        • 0 Attachment
          --- In linux@yahoogroups.com, Scott <scottro@...> wrote:
          >
          > On Tue, May 01, 2012 at 10:57:28AM -0000, thad_floryan wrote:
          > > [...]
          > > I can ssh into the system just fine within 20 seconds or so of
          > > powering up, but the graphic login screen is delayed and there is
          > > no disk activity whatsoever for quite a long time. I didn't time
          > > the number of minutes. Once logged on, I can logout and a new
          > > login window appears almost instantly, so this is something
          > > related to being just-booted.
          >
          > What happens if you boot into runlevel 3 and then do startx. I put
          > nothing past Gnome people or RH in breaking things.

          I haven't explicitly tried that (yet) because X is running fine, it's
          just the presentation of the login window is delayed many minutes.

          We already know that some idiot at Red Hat mucked with the login in
          RHEL/Centos 6* since it worked fine in CentOS 5 and, in fact, has been
          fine since (at least) Red Hat 9 circa 2003. You may remember my bug
          finding last month with all user names revealed in the login window
          which is fixable and what I implemented; RH's argument the fix breaks
          smart card support is specious -- those idiots broke something that
          was working fine.

          I suspect the problem is with one of these:

          /usr/share/gdm/autostart/LoginWindow/gdm-simple-greeter.desktop
          /usr/share/gdm/gdm-greeter-login-window.ui
          /usr/lib64/python2.6/site-packages/orca/scripts/apps/gdmlogin.py
          /usr/lib64/python2.6/site-packages/orca/scripts/apps/gdmlogin.pyc
          /usr/lib64/python2.6/site-packages/orca/scripts/apps/gdmlogin.pyo
          /etc/pam.d/gdm-autologin

          and what's really annoying is seeing nutcase Poettering's grubby
          fingers on this:

          /var/lib/yum/yumdb/p/8baa1b4a5662d830ef508bc8f9c255c5762e524f-pulseaudio-gdm-hooks-0.9.21-13.el6-x86_64

          whatinell is Poettering doing mucking with gdm?


          > > If anyone could reveal the name of the process that puts up the
          > > login window, I could take it from there. SELINUX is *NOT*
          > > enabled as best as I can tell.
          >
          > sestatus is the command to check that.

          Thank you! I didn't know that command. It reports "disabled" which
          is what I selected during the install last month.

          > Also /etc/sysconfig/selinux.

          Thanks again; that's now in my notes for things to check [later].

          Y'know, it's all these kinds of regressive bugs [*ALL* distros] that
          give Linux a bad name and prevent greater adoption on the desktop and
          [often] as servers. Never had these kind of inexcusable problems with
          Solaris, HP-UX, AIX, or any *BSD.

          I've wasted already 5 hours tracking down this bug which mysteriously
          "just" appeared today; logins from cold boots were fine Monday and
          ever since the install on March 11, 2012 (hard to believe it's been
          that long already).
        • Scott
          ... Whatever they let him do. I came to the conclusion, recently, that while we (not only you and I, but a generic we, ) get on poor Mr. Poettering s case,
          Message 4 of 11 , May 1, 2012
          • 0 Attachment
            On Wed, May 02, 2012 at 12:45:03AM -0000, thad_floryan wrote:
            > --- In linux@yahoogroups.com, Scott <scottro@...> wrote:
            > >
            > > On Tue, May 01, 2012 at 10:57:28AM -0000, thad_floryan wrote:
            > > > [...]
            >
            > /var/lib/yum/yumdb/p/8baa1b4a5662d830ef508bc8f9c255c5762e524f-pulseaudio-gdm-hooks-0.9.21-13.el6-x86_64
            >
            > whatinell is Poettering doing mucking with gdm?

            Whatever they let him do. I came to the conclusion, recently, that
            while we (not only you and I, but a generic we, ) get on poor
            Mr. Poettering's case, the fault is with those who feel his ideas
            should go into these systems and put them in without proper testing or
            documentation. It's tempting to blame Mr. Poettering, but he doesn't
            own RH, and I doubt he makes all these decisions himself.

            And, I don't think it's just RH. There is a gradual Windows-ization of
            Linux, hiding things behind GUIs, making it harder and harder to get to
            what controls the GUI.

            On the bright side, we do have more options than with MS or Apple, as
            far as choosing desktop, type of login, and so on.

            >
            > Y'know, it's all these kinds of regressive bugs [*ALL* distros] that
            > give Linux a bad name and prevent greater adoption on the desktop and
            > [often] as servers. Never had these kind of inexcusable problems with
            > Solaris, HP-UX, AIX, or any *BSD.

            My impression, which I think I've given before, is that in the old
            days, most developers were also sysadmins. Now they're probably users
            who grew up thinking Apple is the epitome of what a computer should be.

            And, I've got to admit I'm a bit of a hypocrite. I enjoy being able to
            type yum install or apt-get install, not worrying about dependencies,
            not worrying about compiling my own network card drivers, not needed to
            find out my monitor's exact specs. So, while I rail against the
            catering to the newcomer, I also take advantage of it.


            --
            Scott Robbins
            PGP keyID EB3467D6
            ( 1B48 077D 66F6 9DB0 FDC2 A409 FA54 EB34 67D6 )
            gpg --keyserver pgp.mit.edu --recv-keys EB3467D6

            Buffy: You're missing the whole point of Halloween.
            Willow: Free candy?!
          • thad_floryan
            ... I have read a bunch of interviews with him (and I have the URLs of those on file) and it comes down to the simple fact he simply doesn t give a damn who
            Message 5 of 11 , May 1, 2012
            • 0 Attachment
              --- In linux@yahoogroups.com, Scott <scottro@...> wrote:
              >
              > On Wed, May 02, 2012 at 12:45:03AM -0000, thad_floryan wrote:
              > > --- In linux@yahoogroups.com, Scott <scottro@> wrote:
              > > >
              > > > On Tue, May 01, 2012 at 10:57:28AM -0000, thad_floryan wrote:
              > > > > [...]
              > >
              > > /var/lib/yum/yumdb/p/8baa1b4a5662d830ef508bc8f9c255c5762e524f-pulseaudio-gdm-hooks-0.9.21-13.el6-x86_64
              > >
              > > whatinell is Poettering doing mucking with gdm?
              >
              > Whatever they let him do. I came to the conclusion, recently, that
              > while we (not only you and I, but a generic we, ) get on poor
              > Mr. Poettering's case, the fault is with those who feel his ideas
              > should go into these systems and put them in without proper testing or
              > documentation. It's tempting to blame Mr. Poettering, but he doesn't
              > own RH, and I doubt he makes all these decisions himself.

              I have read a bunch of interviews with him (and I have the URLs of
              those on file) and it comes down to the simple fact he simply doesn't
              give a damn who his changes affect and it's clearly obvious (at least
              to me given what I learned on contract to JPL in the early 1970s to
              establish and prove the tenets of software DESIGN and portability)
              that he has NO concept of software design.

              People who call themselves "software engineers" are blatant liars --
              they have NO concept of engineering disciplines and tools (such as
              schematics for the EE, blueplints for the CEs and MEs, and flowcharts
              for the [alleged] SEs).

              You'd be surprised how many Stanford and UCB grads I refused to hire
              because they couldn't demonstrate any software design concepts, and
              that problem is even worse today even at companies that should know
              better (such as Linuxcare aka Levanta where I ran the IT group and
              had to interface with so-called "software engineers"). At Tymshare I
              was the author of their database system and all engineering software
              (ECAP, MICAP, COGO, etc.) and also the co-author of the first commercial relational DBMS (in the early 1970s) -- I do know what I'm
              doing and 4+ decades of success proves it. :-)

              With that written, I downloaded all the Red Hat 6.2 PDF docs from:

              <http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/index.html>

              and searched all them simultaneously using Adobe Reader (on one of my
              Windows boxes because nothing on Linux will search and index multiple
              PDFs) and found some relevant info regarding PAM and GDM along with
              100s of bugs -- I'm tempted to simply delete /etc/pam.d/ because it
              and GDM don't appear to be enterprise quality. But I'll defer doing
              that until examining gconftool2 and/or gconf-editor which we found by
              the PDF search.

              I'm still really PO'd this problem appeared suddenly today after all
              the prior successful logins. I need to also check the log of all
              installed updates for additional clues since Red Hat clowns seem to
              delight in introducing regressions.


              > And, I don't think it's just RH. There is a gradual Windows-
              > ization of Linux, hiding things behind GUIs, making it harder and
              > harder to get to what controls the GUI.

              Gradual!? You better sit down: up next is Linux's Registry file just
              like Windows. Don't believe? Read the grub2 docs for a starter. :-)

              > On the bright side, we do have more options than with MS or Apple,
              > as far as choosing desktop, type of login, and so on.

              True, but most of the options are superfluous. :-)

              Hey, all I want to do is login and do my work, not wait 15 minutes for
              the login window to appear as it's doing now on my CentOS 6.2 system
              since today.

              > [...]
              > My impression, which I think I've given before, is that in the old
              > days, most developers were also sysadmins.

              True. That was true for me (as a developer) since the sysadmins had
              no clue how the systems operate internally nor could they fix bugs
              so I had to wear multiple hats from 1965 through 2008

              > Now they're probably users who grew up thinking Apple is the
              > epitome of what a computer should be.

              Apple is NOT well-respected here in Silicon Valley though 40 miles
              North of here in 'Frisco it's big among all the hipsters as you can
              see here:

              <http://dbagging.com/level-500-hipster/> :-)

              A friend of mine in Reno took his Mac Pro to the so-called "genius"
              bar today because Apple Mail is a royal POS and email filters set up
              using it prevented mail from me (and others) reaching him. I keep
              recommending that he install Thunderbird which would also permit use
              of GPG. But he's 84 years old and a bit set in his ways. Sigh. The
              problem was "solved" by the Apple "genius" deleting all the filters.
              D'oh, head slap and face palm. :-)

              > And, I've got to admit I'm a bit of a hypocrite. I enjoy being
              > able to type yum install or apt-get install, not worrying about
              > dependencies, not worrying about compiling my own network card
              > drivers, not needed to find out my monitor's exact specs. So,
              > while I rail against the catering to the newcomer, I also take
              > advantage of it.

              No problemo -- there's nothing wrong with a well-designed GUI. A
              number of my products over the years had them and though I'm a
              command-line kind of guy I appreciate a good GUI, too -- saves time.

              If you want to see some really bad GUIs and related IT SNAFUs/FUBARs:

              <http://thedailywtf.com/Articles/>

              :-)
            • thad_floryan
              Following are the verbatim contents of /var/log/gdm/:0-greeter.log.4 written at about the time of the first login failure early this AM which shows clear
              Message 6 of 11 , May 1, 2012
              • 0 Attachment
                Following are the verbatim contents of /var/log/gdm/:0-greeter.log.4
                written at about the time of the first login failure early this AM
                which shows clear evidence of developer incompetence (lines are folded
                around column 65 due to YAHU's bugs). This sure makes me think about
                dumping CentOS and going with FreeBSD -- I don't need this kind of
                crap interfering with my work and projects:

                ** (<unknown>:2300): DEBUG: Client registered with session manager: \
                /org/gnome/SessionManager/Client1
                ** (process:2316): DEBUG: Greeter session pid=2316 display=:0.0 \
                xauthority=/var/run/gdm/auth-for-gdm-coLLfq/database

                (polkit-gnome-authentication-agent-1:2319): GLib-GObject-WARNING \
                **: cannot register existing type `_PolkitError'

                (polkit-gnome-authentication-agent-1:2319): GLib-CRITICAL **: \
                g_once_init_leave: assertion `initialization_value != 0' failed
                Failed to play sound: File or data not found

                gdm-simple-greeter[2316]: Gtk-WARNING: gtkwidget.c:5460: widget \
                not within a GtkWindow

                Window manager warning: Buggy client sent a _NET_ACTIVE_WINDOW \
                message with a timestamp of 0 for 0x180002b (Login Wind)

                Window manager warning: meta_window_activate called by a pager \
                with a 0 timestamp; the pager needs to be fixed.

                Window manager warning: Buggy client sent a _NET_ACTIVE_WINDOW \
                message with a timestamp of 0 for 0x180002b (Login Wind)

                Window manager warning: meta_window_activate called by a pager \
                with a 0 timestamp; the pager needs to be fixed.

                Window manager warning: CurrentTime used to choose focus window; \
                focus window may not be correct.

                Window manager warning: Got a request to focus the no_focus_window \
                with a timestamp of 0. This shouldn't happen!
              • Scott
                ... In that case, who is letting him put these things into Fedora and RedHat. They are the ones to blame. :) -- Scott Robbins PGP keyID EB3467D6 ( 1B48 077D
                Message 7 of 11 , May 2, 2012
                • 0 Attachment
                  On Wed, May 02, 2012 at 04:04:02AM -0000, thad_floryan wrote:
                  > --- In linux@yahoogroups.com, Scott <scottro@...> wrote:
                  > >

                  > > Whatever they let him do. I came to the conclusion, recently, that
                  > > while we (not only you and I, but a generic we, ) get on poor
                  > > Mr. Poettering's case, the fault is with those who feel his ideas
                  > > should go into these systems and put them in without proper testing or
                  > > documentation. It's tempting to blame Mr. Poettering, but he doesn't
                  > > own RH, and I doubt he makes all these decisions himself.
                  >
                  > I have read a bunch of interviews with him (and I have the URLs of
                  > those on file) and it comes down to the simple fact he simply doesn't
                  > give a damn who his changes affect and it's clearly obvious (at least
                  > to me given what I learned on contract to JPL in the early 1970s to
                  > establish and prove the tenets of software DESIGN and portability)
                  > that he has NO concept of software design.

                  In that case, who is letting him put these things into Fedora and
                  RedHat. They are the ones to blame. :)


                  --
                  Scott Robbins
                  PGP keyID EB3467D6
                  ( 1B48 077D 66F6 9DB0 FDC2 A409 FA54 EB34 67D6 )
                  gpg --keyserver pgp.mit.edu --recv-keys EB3467D6

                  Buffy: Could I be seeing Billy's asteroid body?
                  Giles: Astral body, and I don't know.
                • thad_floryan
                  ... Agreed 100%. Seems Red Hat is destined to continue producing only toyware by their prima donna alleged developers . The messages I poster earlier today
                  Message 8 of 11 , May 2, 2012
                  • 0 Attachment
                    --- In linux@yahoogroups.com, Scott <scottro@...> wrote:
                    > [...]
                    > In that case, who is letting him put these things into Fedora and
                    > RedHat. They are the ones to blame. :)

                    Agreed 100%.

                    Seems Red Hat is destined to continue producing only toyware by their
                    prima donna alleged "developers".

                    The messages I poster earlier today from /var/log/gdm/:0-greeter.log.4
                    are a clear sign Red Hat has NO QA procedures in place and RH software
                    is not to be trusted. Logins used to work fine and they regressed it
                    as can be seen here:

                    <https://bugzilla.redhat.com/show_bug.cgi?id=666220>
                    and
                    <http://tech.groups.yahoo.com/group/linux/message/60489>

                    Mike Grima (at Red Hat) claims they won't fix their new and regressed
                    login problems introduced in 6.0 until 6.3 or 6.4.

                    That's ridiculous; their present (stock) login presentation is a
                    security mess with all valid user names displayed on the screen
                    whenever the login window does appear.

                    That's it. It's clear Red Hat does NOT produce enterprise quality
                    software.

                    *P*O*O*F*, CentOS will be history here later today after I download
                    and install/replace it with a real OS. I wasted 6 weeks of my time
                    "playing" with a toy POS non-enterprise-level distro since March 11,
                    2012. Speaking of which, they still never fixed the Live CD bug I
                    reported in 2008:

                    <http://bugs.centos.org/view.php?id=3123>

                    and mentioned here again on April 11, 2010 (2+ years ago):

                    <http://tech.groups.yahoo.com/group/linux/message/57300>

                    Seeing how they (Red Hat & CentOS) support their stuff, anyone using
                    their software may want to rethink using it for any real tasks in the
                    business world.

                    Doesn't leave many choices: Arch Linux is one, and the *BSDs are
                    another, and I favor FreeBSD which is used on my website and has a
                    demonstrated history of reliability.
                  • Scott
                    ... There was one case where Alan Bartlett of CentOS (not officially, but very active with them, helps maintain the Elrepo packages of device drivers), filed a
                    Message 9 of 11 , May 2, 2012
                    • 0 Attachment
                      On Wed, May 02, 2012 at 05:53:41PM -0000, thad_floryan wrote:
                      > --- In linux@yahoogroups.com, Scott <scottro@...> wrote:


                      >
                      > Seems Red Hat is destined to continue producing only toyware by their
                      > prima donna alleged "developers".
                      >
                      > The messages I poster earlier today from /var/log/gdm/:0-greeter.log.4
                      > are a clear sign Red Hat has NO QA procedures in place and RH software
                      > is not to be trusted. Logins used to work fine and they regressed it
                      > as can be seen here:

                      >
                      > <https://bugzilla.redhat.com/show_bug.cgi?id=666220>

                      There was one case where Alan Bartlett of CentOS (not officially, but very active with them, helps maintain the
                      Elrepo packages of device drivers), filed a report on that. IIRC, their attitude was that you can edit the file to fix it.


                      > and
                      > <http://tech.groups.yahoo.com/group/linux/message/60489>
                      >
                      > Mike Grima (at Red Hat) claims they won't fix their new and regressed
                      > login problems introduced in 6.0 until 6.3 or 6.4.

                      There are a lot of decisions they made that seem, to me, to be oriented only towards the desktop--not only that one, but
                      crippling the text install, and so on. They recently removed their system-config-bind thingie with no replacement save manual
                      configuration. And so on. I keep thinking it's a good chance for Canonical, especially as there are some old timers from the days
                      when Debian was cooler than RH.


                      > Doesn't leave many choices: Arch Linux is one, and the *BSDs are
                      > another, and I favor FreeBSD which is used on my website and has a
                      > demonstrated history of reliability.

                      Jeff has pointed out that in many ways, unfortunately, both of those are some niche distributions--well, the BSDs aren't
                      distributions, but you know what I mean. One reason an individual can get a really quick response is their relatively small numbers
                      of users. This is good, of course, in many ways, but it's the RedHats and Canonicals who get us the driver support, fight the
                      legal battles, hire Jeff (they both did, though, AFIK, not simultaneously), and so on.


                      --
                      Scott Robbins
                      PGP keyID EB3467D6
                      ( 1B48 077D 66F6 9DB0 FDC2 A409 FA54 EB34 67D6 )
                      gpg --keyserver pgp.mit.edu --recv-keys EB3467D6

                      Buffy: Well, that works out great. You won't tell anyone
                      that I'm the Slayer, and I won't tell anyone you're a moron.
                    • thad_floryan
                      ... I ve done a lot of such fixes for a number of things (Perl, Emacs, etc.) but this new mess seems possibly attributed to the kernel [re: pager]. Look at
                      Message 10 of 11 , May 2, 2012
                      • 0 Attachment
                        --- In linux@yahoogroups.com, Scott <scottro@...> wrote:
                        >
                        > On Wed, May 02, 2012 at 05:53:41PM -0000, thad_floryan wrote:
                        > > [...]
                        > > <https://bugzilla.redhat.com/show_bug.cgi?id=666220>
                        >
                        > There was one case where Alan Bartlett of CentOS (not officially, but
                        > very active with them, helps maintain the Elrepo packages of device
                        > drivers), filed a report on that. IIRC, their attitude was that you
                        > can edit the file to fix it.
                        > [...]

                        I've done a lot of such fixes for a number of things (Perl, Emacs, etc.)
                        but this new mess seems possibly attributed to the kernel [re: pager].

                        Look at these message again -- there's NO WAY I'm going to waste my time
                        fixing something that shouldn't have been broken in the first place
                        (and note I folded the lines circa column 65 with a "\" due to nasty
                        YAHU bugs); these messages from /var/log/gdm/:0-greeter.log.4:


                        ** (<unknown>:2300): DEBUG: Client registered with session manager: \
                        /org/gnome/SessionManager/Client1
                        ** (process:2316): DEBUG: Greeter session pid=2316 display=:0.0 \
                        xauthority=/var/run/gdm/auth-for-gdm-coLLfq/database

                        (polkit-gnome-authentication-agent-1:2319): GLib-GObject-WARNING \
                        **: cannot register existing type `_PolkitError'

                        (polkit-gnome-authentication-agent-1:2319): GLib-CRITICAL **: \
                        g_once_init_leave: assertion `initialization_value != 0' failed
                        Failed to play sound: File or data not found

                        gdm-simple-greeter[2316]: Gtk-WARNING: gtkwidget.c:5460: widget \
                        not within a GtkWindow

                        Window manager warning: Buggy client sent a _NET_ACTIVE_WINDOW \
                        message with a timestamp of 0 for 0x180002b (Login Wind)

                        Window manager warning: meta_window_activate called by a pager \
                        with a 0 timestamp; the pager needs to be fixed.

                        Window manager warning: Buggy client sent a _NET_ACTIVE_WINDOW \
                        message with a timestamp of 0 for 0x180002b (Login Wind)

                        Window manager warning: meta_window_activate called by a pager \
                        with a 0 timestamp; the pager needs to be fixed.

                        Window manager warning: CurrentTime used to choose focus window; \
                        focus window may not be correct.

                        Window manager warning: Got a request to focus the no_focus_window \
                        with a timestamp of 0. This shouldn't happen!
                      • thad_floryan
                        This is almost unbelievable -- I fixed the CentOS Gnome login FUBAR. I was reading some FreeBSD docs in preparation to blow-away CentOS and replace it with
                        Message 11 of 11 , May 2, 2012
                        • 0 Attachment
                          This is almost unbelievable -- I "fixed" the CentOS Gnome login FUBAR.

                          I was reading some FreeBSD docs in preparation to blow-away CentOS and
                          replace it with FreeBSD when I came across this link:

                          <http://www.freebsd.org/gnome/docs/faq2.html#q20>

                          which has the following lines copy'n'pasted:
                          "
                          " [...]
                          " 20. Why is GNOME so slow to start up?
                          "
                          " Under normal circumstances, GNOME should only take a few seconds
                          " to start up (fifteen at most). However, certain configurations
                          " may cause it to hang for up to an hour at login time.
                          "
                          " First, make sure your machine's hostname properly resolves. To
                          " test this, run the following command:
                          "
                          " ping `hostname`
                          "
                          " If the command fails, you will either have to add your
                          " fully-qualified hostname to DNS or to /etc/hosts. If you do not
                          " have a static IP address, you can append your hostname to the
                          " localhost line in /etc/hosts.
                          " [...]

                          Which got me wondering since I posted my makehosts script the day
                          before the login problem began and copied the results to /etc/hosts
                          on the CentOS box because it only had a minimal hosts file and that
                          was the only change I had made to the system in awhile.

                          The suggested "ping `hostname`" worked fine because I had this line
                          in /etc/hosts

                          172.20.20.63 procyon

                          I commented that line out per "#172.20.20.63 procyon", rebooted, and
                          the login window came up fine after a reboot.

                          WTF?!

                          So I added the hostname to the localhost line ending up with this:

                          127.0.0.1 localhost procyon localhost.localdomain
                          [...]
                          #172.20.20.63 procyon
                          [...]

                          then rebooted and the login window came up fine again.

                          Needless to say, this is bizarre. The above discovery has caused me
                          to not blow-away CentOS at least for the time being :-)

                          The symptoms suggest there's some really flaky network coding in Gnome
                          for remote graphical logins so my earlier comments about incompetent
                          developers (RedHat, Gnome, et al) is corroborated. :-)
                        Your message has been successfully submitted and would be delivered to recipients shortly.